The gcc environment is installed on the system (otherwise, compiling redis will cause errors).Still don't know how to install the gcc environment offline viewable Install gcc environment offline under CentOS with detailed graphics

download

Click on the official address to download redis for Linux

http://download.redis.io/releases/redis-5.0.4.tar.gz

upload

put D:\AppStore\redis\redis-5.0.4.tar.gz  /root

install

tar xzf redis-5.0.4.tar.gz
mv redis-5.0.4  /usr/local/redis
cd /usr/local/redis/redis-5.0.4
make
make install

If make errors occur as follows

Fatal error: jemalloc/jemalloc.h: No file or directory

Then use

make MALLOC=libc

Node Configuration

Now you have three virtual machines, you can do it all the time.Enter redis.conf and change as follows

192.168.1.115 (primary node)

bind 192.168.1.115
daemonize yes
masterauth 666666
requirepass 666666

pidfile /st/redis/pid/redis_6379.pid
logfile "/st/redis/log/redis.log"
dir /st/redis/db/

192.168.1.113 (from Node 1)

bind 192.168.1.113
daemonize yes
masterauth 666666
requirepass 666666
replicaof 192.168.1.115 6379

pidfile /st/redis/pid/redis_6379.pid
logfile "/st/redis/log/redis.log"
dir /st/redis/db/

192.168.1.112 (from Node 2)

bind 192.168.1.112
daemonize yes
masterauth 666666
requirepass 666666
replicaof 192.168.1.115 6379

pidfile /st/redis/pid/redis_6379.pid
logfile "/st/redis/log/redis.log"
dir /st/redis/db/

Sentry Configuration

Now that you have three virtual machines, don't waste it; just pair three sentries on them.However, if a server is hung, the node will be hung, and the sentry on that server will be hung, but there are still two senders who can work, so it is acceptable.Otherwise, set up separate servers for the sentry.

Enter sentinel.conf and change as follows

192.168.1.115

daemonize yes
sentinel monitor mymaster 192.168.1.115 6379 2
sentinel auth-pass mymaster 666666

pidfile /st/redis/pid/redis-sentinel.pid
logfile "/st/redis/log/redis-sentinel.log"
dir /st/redis/tmp

192.168.1.113

daemonize yes
sentinel monitor mymaster 192.168.1.115 6379 2
sentinel auth-pass mymaster 666666

pidfile /st/redis/pid/redis-sentinel.pid
logfile "/st/redis/log/redis-sentinel.log"
dir /st/redis/tmp

192.168.1.112

daemonize yes
sentinel monitor mymaster 192.168.1.115 6379 2
sentinel auth-pass mymaster 666666

pidfile /st/redis/pid/redis-sentinel.pid
logfile "/st/redis/log/redis-sentinel.log"
dir /st/redis/tmp

Start Node

From primary to secondary, each node runs and checks

/usr/local/redis/redis-5.0.4/src/redis-server /usr/local/redis/redis-5.0.4/redis.conf

netstat -tulnp |grep redis

Start Sentinel

Each Sentinel runs and checks

/usr/local/redis/redis-5.0.4/src/redis-sentinel /usr/local/redis/redis-5.0.4/sentinel.conf

ps -ef |grep redis

inspect

Check master-slave replication, log in to the master node to write information and see if the information is received from the node

/usr/local/redis/redis-5.0.4/src/redis-cli -h 192.168.1.115 
auth 666666

set name 'lgx'

Log on from the node to see if you can receive information

/usr/local/redis/redis-5.0.4/src/redis-cli -h 192.168.1.113 
auth 666666

get name 'lgx'

Check Sentinel Monitoring

/usr/local/redis/redis-5.0.4/src/redis-sentinel -p 26379

sentinel master mymaster

Register a global directory if you dislike using the previous heap of commands

ln -s /usr/local/redis/redis-5.0.4/src/redis-sentinel /usr/bin/redis-sentinel
ln -s /usr/local/redis/redis-5.0.4/src/redis-server /usr/bin/redis-server
ln -s /usr/local/redis/redis-5.0.4/src/redis-cli /usr/bin/redis-cli

Matters needing attention

On production remember to check: firewall, security group, redis password, bound ip, storage directory of each file.

This is the default directory to avoid forgetting the configuration and not knowing where to put it.

#redis node
pidfile /var/run/redis_6379.pid
logfile ""
dbfilename dump.rdb
dir ./

#redis Sentinel
pidfile /var/run/redis-sentinel.pid
logfile ""
dir /tmp

This is adding node ports, sentry ports to the firewall

firewall-cmd --permanent --add-port=6379/tcp
firewall-cmd --permanent --add-port=26379/tcp//Add a port

firewall-cmd --reload  //Then reload

Test Report

When the primary node is down, it restarts immediately. It takes some time, about a few seconds, for the primary node to switch. During this time, if you operate redis, the operation will fail because the connection to the primary node is unsuccessful.

When switching, the configuration of the node and the sentry is automatically modified, such as the replicaof parameter from the node and sentinel monitor mymaster parameter from the sentry.

After the downtime restarts, nodes and sentries cannot recover automatically and need to be started manually.I thought daemonize yes means start-up and start-up. It could have been run in the background or exit the redis command line page.